Analytical Applications of the Electrochemical Quartz Crystal

Apr 23, 1992 - A. Robert Hillman1, David C. Loveday1, Marcus J. Swann1, Stanley Bruckenstein2, and C. Paul Wilde3. 1 School of Chemistry, University o...
0 downloads 0 Views 1MB Size
Chapter 12

Analytical Applications of the Electrochemical Quartz Crystal Microbalance 1

1

1

A. Robert Hillman , David C . Loveday , Marcus J. Swann , Stanley Bruckenstein , and C . Paul Wilde 2

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

1

3

School of Chemistry, University of Bristol, Bristol BS8 1TS, England Department of Chemistry, State University of New York, Buffalo, N Y 14214 Department of Chemistry, University of Ottawa, Ottawa, Ontario K 1 N 6N5, Canada 2

3

The electrochemical quartz crystal microbalance ( E Q C M ) has considerable potential as a sensor. The strategy is to use exchange of mobile species between a surface immobilised polymer film and a solution as a probe of solution composition. Film composition changes are monitored gravimetrically. The quartz crystal microbalance offers generality of detection by mass and high sensitivity. Selectivity is sought via choice of polymer coating and electrochemical control variables. In this paper we discuss the thermodynamics and kinetics of the exchange process and conditions under which the EQCM technique is applicable. Background Modification of electrode surfaces with polymer films has been the subject of considerable recent research (1). Potential areas of application include electrical/optical devices, surface protection, energy conversion/storage, electrocatalysis and electroanalysis. In this paper we discuss issues relevant to the use of polymer modified electrodes in sensors. The strategy upon which we focus here involves the exchange of mobile species (the target species and, possibly, interférants) between a polymer film (the sensor) and its bathing solution (the analyte). Extraction based on ion-exchange into a charged polymer film has been exploited for determination of metal species (2),(3) and neurotransmitters (4). Complexation by an appropriately chosen ligand is also possible (5), and the specificity of antibody interactions has been exploited (6). Following the immobilisation of the target species on the electrode, voltammetric determination of electroactive ions is straightforward. Although electroinactive ions may be determined indirectly, via their effect on the competitive binding of electroactive ions (7), a more direct approach is clearly preferable. Uptake of solution species by the surface-modifying polymer film offers 0097-6156/92/0487-Ο150$06.00/0 © 1992 American Chemical Society

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

12. HILLMAN ET AL.

Electrochemical Quartz Crystal Microbalance

151

the opportunity for pre-concentration. The extent to which this may enhance sensitivity may be determined by both thermodynamic and kinetic factors. When thermodynamic effects predominate (if the system is allowed to come to equilibrium), activity effects are significant (8). Equality of activity of mobile species will, in general, result in the transfer of neutral species, notably solvent, between the solution and polymer phases. Furthermore, activity effects may contribute to the observed (2) non-linearity of concentration-based isotherms for mobile species partition into the polymer film. Kinetic effects may be superimposed on the thermodynamic requirements. A t first sight, slow solution / polymer exchange is expected to be deleterious, since it lowers sensitivity (2),(3),(9). However, we demonstrate here the principles of a method by which differences in mobile species transfer rates may be turned to advantage. Selectivity is sought by appropriate choice of the polymer, based on established chemical principles. For ion-exchange, there is a crude selectivity based on charge-type (9), provided the film is permselective (see below). Additionally, some selectivity between ions of similar charge type may be based on charge number. For complexation, the use of selective ligands is common in analytical chemistry. In practice, more than one type of interaction may contribute to the selectivity pattern (2)(4). The Electrochemical Quartz Crystal Microbalance ( E Q C M ) The resonant frequency of a quartz crystal oscillator is perturbed from its base value (Îq) by attached overlayers. For thin, rigid films the measured change in resonant frequency (Af) with attached mass (ΔΜ) is described by the Sauerbrey equation (10): 2

2

(Af/Hz) = - (2/pv)fo (AM/g cm" )

[1]

where ρ is the density of the quartz and ν is the wave velocity. The quartz crystal microbalance ( Q C M ) technique has been used for many years (11) to monitor the deposition of materials (e.g. metals) from the vapour phase onto solid substrates. More recently, it was shown that the crystal continues to oscillate when one face is exposed to a liquid (12),(13). In the E Q C M , the electrode exposed to the solution is used as the working electrode in an electrochemical cell. The capabilities of the E Q C M have been reviewed recently (14),(15),(16). Of primary concern here is the ability to monitor the exchange of mobile species between a polymer film and its bathing electrolyte (17), (18), (19), (20), (21), (22), (23), (24), (25), (26), (27), (28), (29). The E Q C M can detect overlayer formation, dissolution or composition changes. When these can be related to the composition of the ambient medium, this provides the basis of a sensor. Key advantages are the generality of detection by mass and the high sensitivity of the E Q C M : in situ mass changes of 1 ng cm" can routinely be detected. Several applications of the Q C M to sensing molecules of biological importance have been described. The adsorption of cholesterol in thin films on a Q C M has been used as a model for biological 2

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

152

BIOSENSORS AND CHEMICAL SENSORS

chemoreception in lipid bilayers (30), and the question of partition coefficient considered. Chemical amplification of the Q C M sensitivity has been neatly demonstrated using antibody reactions (6). Another interesting example involved glucose detection via binding to hexokinase in a polyacrylamide film on a Q C M (31). Here the sensitivity was markedly greater than predicted on the basis of equation [1], a phenomenon ascribed to changes in polymer Theological properties consequent upon uptake of the target species. In this paper we discuss three issues related to our ability to exploit the undoubted attractions of the E Q C M technique: (a) the extent of mobile species uptake as a function of solution concentration; (b) the use of transient measurements to obtain (additional) selectivity and (c) the need to establish that the criteria are satisfied for the Sauerbrey equation (equation [1]) to be used to convert measured frequency changes to mass changes. Of these, (a) and (c) have been demonstrated (see previous paragraph) to be directly relevant to QCM-based biosensors. The concept of using transient measurements in this context has not yet been explored, but is a natural development. Experimental The instrumentation (24)-(26) and the general technique (12) have been described previously. Polythionine (PTh) films were deposited by the electrochemical polymerisation (at 1.1V) of thionine from aqueous solutions of 0.05 mol dm"* H C 1 0 (24). Polyvinylferrocene ( P V F ) films were electrodeposited (at 0.7V) from solutions of P V F in C H C 1 / 0.1 mol d m tetrabutylammonium perchlorate (TBAP) (30). Polybithiophene (PBT) films were deposited by the electrochemical polymerisation (at 1.225V) of 2,2'bithiophene in C H C N / 0.1 mol dm" tetraethylammonium tetrafluoroborate ( T E A T ) (26). After deposition, all modified electrodes were transferred to monomer/polymer-free solutions (see figure legends for composition) for characterisation. The electrodes (area 0.23 cm ) on the quartz crystals were A u for PTh and P V F , and Pt for PBT. Polymer coverages (reported in terms of moles of electroactive sites, Γ/mol cm* ) were determined by integration of slow scan rate (< 5 m V s" ) voltammetric current responses. For different films, these were typically in the ranges 3-10 nmol c m (PTh), 5-12 nmol cm" (PVF), and 16-38 nmol cm" (PBT). Experimental values of mass changes, ΔΜ / ng cm* , are normalised by multiplying by the Faraday and dividing by the charge passed, Q / μC cm" . The resulting "normalised mass change", A M F / Q , then corresponds to the mass change associated with redox switching of one mole of redox sites. This facilitates comparison of data for different films (different polymer coverages). 4

-3

2

2

3

3

2

2

1

- 2

2

2

2

2

Results and Discussion Thermodynamics of mobile species uptake. Operationally, one seeks a linear, or at least single valued, relationship between the film mass change and analyte composition. In this section we stress the importance of characterising the mass change / composition relationship, and illustrate circumstances under which the desired behaviour will not prevail.

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

12. HILLMAN ET AL.

153

Electrochemical Quartz Crystal Microbalance

Activity effects. The exchange of trace ions in solution with others in the polymer film might, simplistically, be expected to lead to a linear uptake/solution concentration relationship. Unfortunately, this is seldom the case. The thermodynamic restraint is that of electrochemical potential. Thus electroneutrality is not the sole constraint on the ion exchange process. A second thermodynamic requirement is that the activity of mobile species in the polymer and solution phases be equal. (Temporal satisfaction of these two constraints is discussed below, with reference to Figure 4.) The rather unusual, high concentration environment in the polymer film can lead to significant - and unanticipated - activity effects (8). This has been demonstrated in E Q C M studies of P T h film redox switching in H C 1 0 solutions of different concentration. Under conditions where thionine reduction is a 2 e / 3 H process, electroneutrality alone predicts uptake of one anion (and three protons) per T h site: a film mass increase of 102.5 g molTh" , independent of solution composition. Experimentally, the mass change is less than 20 g m o l ' in 1 mol d m H C 1 0 , decreases as the electrolyte is diluted, and even becomes negative at p H > 2! The variation of mass change with concentration is attributable to activity effects. Hydronium perchlorate is included within the film, to an extent dependent on polymer redox state and solution concentration. 4

+

1

1

- 3

4

Co-ordination effects. Analogous experiments in weak acid solutions (this time at constant total acid concentration, varying p H by additions of base) yield quite different results, as illustrated in Figure 1. The "simple" prediction of electroneutrality is a mass gain of 62 g m o l ' (CH C0 "H-3H ") independent of solution composition. The experimental result is a mass loss: 16 g mol" , independent of anion (acetate) concentration, for p H < 5, decreasing somewhat in magnitude at higher p H . Gas phase experiments, involving exposure of PTh films to dry N , water saturated N and C H C 0 H saturated N show reversible uptake / loss of water, but irreversible uptake of acetic acid. The strength of the interaction between the undissociated acid and the polymer is the key to the behaviour in Figure 1. A t low p H , one (or more) co-ordination sites on each oxidised thionine redox site are occupied by an acetic acid molecule. Upon polymer redox site switching, one H A per redox site dissociates, totally satisfying the demand for counter ion, and partially satisfying the demand for protons. In short, there is no counter ion ingress, because it was present in the film all the time. The remaining two protons required by the reaction are supplied from solution. When their entry is taken into account, the net mass loss of 16 g mol* indicates loss of a species of mass 18. Water is the obvious candidate. We therefore propose that the half-reaction in low p H acetic acid solutions is: 1

:l

3

2

1

2

2

3

2

2

1

+

[ ( T H A - ) · ( H 0 ) · X · ( H A ) ] + 2e + 2 H 0 2

p

[(TH

2 + 4

3

+ S

(A-) ) · X ] + 3H O 2

p

2

= s

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

[2]

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

154

BIOSENSORS AND CHEMICAL SENSORS

where X may be either acetic acid or water, and the number of X's is determined by the maximum coordination number. When the solution p H approaches the p l ^ for acetic acid, the supply of H A within the solution, and thus partitioned into the film, decreases. Consequently, the counter ion must increasingly be supplied from solution, and the mass change moves in a positive direction. The key result in terms of a sensor is that specific interactions, as sought for biochemical sensors (6), may be sufficiently strong that a coordination-type model applies. Note that this does not contradict the activity arguments of the previous section, but is a special case within the general thermodynamic framework. Under these special circumstances, the polymer will be "saturated" with the target species, and film composition will not depend on solution concentration, except at a very low level. Permselectivity: real and apparent Permselectivity is a key issue when the extraction process is based on ion-exchange (31). A straightforward E Q C M illustration is provided by the behaviour of P V F films upon redox switching in aqueous N a C 1 0 solutions (22). A t electrolyte concentrations below 1 mol dm" , the normalised mass change is independent of composition, and indicates ion and solvent entry upon oxidation of neutral P V F . A t higher electrolyte concentrations, permselectivity fails and salt also enters the film upon oxidation. The interpretation of the behaviour of P B T is more subtle. Overall mass changes upon total P B T oxidation / reduction are similar to the counter ion ("dopant") molar mass, for example F A M / Q = 93 g mol" in 0.01-0.1 mol dm" E t N B F 7 C H C N compared to m g p — = 87 g mol" . These results apparently imply permselectivity with little or r\o solvent transfer at low electrolyte concentration, and permselectivity failure at high electrolyte concentration. As we show in the next section, this apparent permselectivity is entirely fortuitous, and results from a compensating combination of mobile species transfers. The message here is that a combination of thermodynamic and kinetic data is required to unequivocally attribute the mass change to the relevant species' transfers. 4

3

1

+

4

3

1

4

3

Kinetics of mobile species uptake. In the previous section, we considered the overall mass changes associated with complete oxidation / reduction of a film, which was then allowed to come to equilibrium with the solution phase. In this section we consider the time course of the transformation, i.e. the transfer of species during the oxidation / reduction. As is common in kinetic studies, it is often more convenient to consider fluxes, rather than populations. For this purpose, we will consider current, i , and the time differential of the mass, $Λ , as well as charge, Q, and mass change, ΔΜ. In analyzing kinetic E Q C M data, we have found it convenient (32)(33) to define a function §>: $>j = M + i (mj/ZjF)

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

P]

12. HILLMAN ET A L

Electrochemical Quartz Crystal Microbalance

155

§>j is defined for a particular ion j , of molar mass nL and charge Z: (including sign). It is a weighted sum of the mass and charge fluxes, with units of g cm" s . The definition of $>j is such that the contribution of the j-th ion's mass is zero. We are free to choose j as any of the ions present, enabling us to systematically eliminate the contribution of each ion to the mass flux. This mathematical device is related to selectivity, which requires determination of changes in individual species populations in the film. A more general discussion of the significance of and the analogous integrated function j (/g cm" ), defined in terms of the mass change and charge, is presented elsewhere (32)(33). Here we make two observations on the nature of Φ: before exploiting it to interpret kinetic data. Firstly, we recognise that mobile species can be divided into charged and net neutral species. The net neutrals contribute to the mass, but not charge, response. The Φ function provides a means of separating ion and neutral species transfers. In the case of a film immersed in a single electrolyte, j (fo) represents the population change (flux) of neutral species (salt and/or solvent;. Secondly, if a given j-th ion is the only mobile species (permselectivity without solvent transfer), j is zero throughout the redox transformation. Furthermore, §>j must also be zero throughout the transformation. The latter is a much more stringent test, since it can detect compensatory transfers of species with different mobility. 2

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

2

Apparent permselectivity and compensatory motion. Although normalised mass change data for the "doping/undoping" of P B T films were very similar to those predicted by permselectivity in the absence of solvent transfer (cf. electroneutrality), the differences were real. Furthermore, systematic variation of the anion or cation or deuteration of the solvent produced consistent trends in the departure from "simple" behaviour. Insight into the overall processes involved (the thermodynamics) is gained by considering the kinetics of mobile species transfer. One way in which this can be done is by considering the shape of the ΔΜ vs. Ε and ΔΜ vs. Q curves, rather than the overall changes in ΔΜ and Q. For a permselective film, electroneutrality requires 1:1 correlation between the electron and counter ion (here, anion) fluxes. If no solvent is transferred, this demands that the ΔΜ vs. Q plot be linear and free of hysteresis, regardless of the shape of the ΔΜ vs. Ε plot. The slope of this plot will be F / m ^ - , where mA ' is the counter ion molar mass. Here we illustrate the utility of the 5>j function defined in the previous section. By "correcting o f f the counter ion contribution to the mass flux (see the form of equation [3]), neutral species fluxes are highlighted: they are the departure of $ j from zero. Experimental mass changes during P B T doping/undoping do not conform to the §>j = 0 requirement. Figure 2 contains representative data for tetraethylammonium hexafluorophosphate ( T E A P F ) as the electrolyte. In accord with the general activity constraint (see above), solvent and salt do transfer. For PBT, these net neutral species transfers are in opposite directions.

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

156

BIOSENSORS AND CHEMICAL SENSORS

4

pH

5

-4 *Q -8

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

^

42

•fe -16 01M CHjCOOH

-20

Figure 1: Normalised mass change for a PTh film in Q i ^ C C ^ H solutions as a function of p H at fixed total acetate concentration (0.1 mol dm ) . (Reproduced from ref. 25. Copyright 1990 American Chemical Society.) â

-0.2 -0.4

0.6

0.8

1

1.2

1.4

1.8

2

Q / mC cm"

Figure 2: Chronoamperometric and chronogravimetric data for a P B T film immersed in 0.1 mol d m ' T E A P F . The potential was stepped from 0V to 1.05V held at 1.05V for 15s, then stepped back to 0V. Values of $ P F ~ were calculated according to equation [3]. Charge data are referred to tne initial condition at 0V. 3

6

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

12. HILLMAN ET AK

Electrochemical Quartz Crystal Microbalance

157

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

The magnitudes of these opposing component mass changes can be rather similar. This fortuitous balance gives the illusion of permselectivity if one considers only the overall mass change. Note that the data of Figure 2 suggest salt exit is faster than solvent entry (during oxidation) and that salt entry is faster than solvent exit (during reduction). In fact, this may indicate (partial) maintenance of electroneutrality by co-ion on short time scales following a potential step. The key results of this section are twofold. First, kinetic data can reveal otherwise hidden thermodynamic information. Second, time can be used as a variable to probe the transfers of individual species, even if several are transferred. This latter aspect is explored in the following section. Kinetic permselectivity. When permselectivity is not achieved in a thermodynamic sense (for example at high electrolyte concentration), we propose that it may be possible to achieve it kinetically. We aim to exploit the differing rates of mobile species transfer. In a transient experiment the response on a short time scale will be dominated by the fastest moving species. The converse will be true on a long time scale. We suggest that this approach might be exploited at two levels. Firstly, field assistance of ion transfers (migration) will lead to their being more rapid than neutral species transfers. Secondly, size effects (for ions or neutral species) will lead to a diversity of transport rates. These effects are likely to be more pronounced in the confined geometry of polymer films than for the same species in solution. The extent to which transfer of a given species dominates the net transfer process (on a given time scale) will depend on its availability, i.e. solution concentration. We first illustrated this type of effect during rapid scan voltammetry of P V F films in concentrated N a C 1 0 solutions, where the overall redox switching process involves ingress of counter ion, salt and solvent upon oxidation (34). Quantitative treatment of such effects is better explored using a potential step, i.e. chronoamperometry. E Q C M data from such an experiment are shown in Figure 3. For comparison purposes, data for analogous experiments in 0.1 and 3 mol dm" electrolyte, where the polymer is / is not permselective, are superimposed. The data in the first frame are the (normalised) current, representing the electron flux. In the second frame, the overall mass flux (transfer of chargecompensating ions as well as neutral species) is shown. The third frame corresponds to the difference between the first two. The quantity &QO -> defined by equation [3] with m; = 99.5 g m o l ' and Ζ: = -1, represents the flux of all species other than perchlorate. In the case o f the dilute (concentrated) electrolyte, this corresponds to the mass flux of water (water+salt). The presence of kinetic permselectivity is demonstrated by comparison of the mass and charge fluxes, in Figure 3. In each case (either electrolyte, either direction of change), the initial slope of the mass flux / current plot corresponds closely to that anticipated for transfer of one counter ion (no salt or solvent) per electron transferred (dashed lines have slope F/99.5). Transfers of the net neutral species, salt and solvent, are purely diffusive and only contribute significantly to the E Q C M response at longer times. 4

3

4

1

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

158

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

BIOSENSORS AND CHEMICAL SENSORS

7000 6000 5000-

-10004.5

5

5.5

6

6.5

t/s

τ , 3000I 2000

t/s

Figure 3: Chronoamperometric data (first frame) and chronogravimetric data (second frame) for a P V F film immersed in 0.1 ( H ) and 3.0 (+) mol d m ' NaC10 . The potential was stepped from 0 V to 0.7V at t=5s. Values of $ciO/~ ( ^ i r d frame) were calculated according to equation [3]. (Adaptea from reference (33), with permission.) 3

4

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

12. HILLMAN ET A L

159

Electrochemical Quartz Crystal Microbalance

A n extreme case of differing transfer rates is anticipated for proton transfer in aqueous media. This is neatly illustrated by E Q C M data for the oxidation of PTh in strong acid media at pH2 (see above). Here the overall mass change is zero, as a result of compensatory motion of ions, acid and solvent. However, during a transient experiment, protons will dominate electroneutrality maintenance on short time scales, while anions (C10 ") will have the opportunity to participate on longer time scales. The difference in tranport rates of the mobile species is graphically demonstrated (35) by the transient mass excursion: a rapid mass increase, followed by a slower decrease as the slower-moving species transfer to satisfy the ultimate requirements of thermodynamics. Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

4

Non-rigidity. A l l the foregoing E Q C M data analyses employed equation [1] to convert measured frequency changes (Af) to mass changes (ΔΜ). This procedure is appropriate for rigid films. If the polymer is extensively solventswollen, this requirement may not be satisfied (36), in which case the simple analytical utility of the E Q C M technique is invalidated. The importance of establishing rigidity in the medium and under the conditions employed cannot therefore be overemphasised. We now present data for ferrocene-based polymer films, showing how rigidity may be a function of both polymer and solution compositions. Exposure to a liquid of one of the electrodes (bare A u or Pt) of the E Q C M results in the establishment of a liquid modulation layer. The thickness of this layer is (12)

where v is the kinematic viscosity of the liquid. For 10 M H z AT-cut crystals exposed to aqueous solutions, x is of the order of 300nm. Effectively, the E Q C M "weighs" this layer of liquid, and the resonant frequency of the crystal decreases, to an extent dependent on electrolyte composition. (Typical frequency changes in aqueous media are 6-7kHz, although additional contributions due to entrapped liquid in surface roughness may raise this number by several kilohertz.) Frequency changes due to the deposition of a rigid film are then simply additive. In the event that the film is non-rigid, this is no longer the case. The key question is "How does one establish film rigidity?" One approach involves analysis of the shape, as well as position, of the resonance. Essentially, one determines the Q-factor for the crystal: broadening signals nonrigidity. It is important to appreciate that immersion of the crystal in the liquid significantly broadens the resonance. Small changes in Q-factor associated with incomplete polymer film rigidity may therefore be difficult to detect, and certainly complex to quantify. A n alternative approach is to compare the frequency data with electrochemical (coulometric) data. This is exemplified for P V F and a co­ polymer with vinylpyrrolidone, poly(vinylferrocene-co-vinylpyrrolidone) (20:80 PVF-co-PVP). Table I shows in and ex situ frequency data for an electrode before and after coating with 20:80 PVF-co-PVP. L

L

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

160

BIOSENSORS AND CHEMICAL SENSORS

Table I. E Q C M data for 20:80 PVF-co-PVP in various environments

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

Experiment

M /Hz

(Af -M) /Hz

ΔΜ

0

a

b

Uncoated, dry crystal in air

-1050

Uncoated crystal in deposition solution

-8025

-6975

7.673

Coated crystal, oxidised film in deposition solution

-11,234

-10,184

11.202

Coated crystal, oxidised film, in air

-14,367

-13,317

14.649

Frequency difference vs. reference crystal. M refers to uncoated, dry crystal. Apparent mass change, assuming rigidity (equation [1]). Entries 2 - 4 refer to liquid modulation layer, liquid modulation layer+polymer and polymer, respectively. 0

The data of column 4 show the apparently paradoxical result that the mass of polymer alone (measured in air) is greater than that of the polymer plus liquid modulation layer. This is clear evidence that the oxidized PVF-co-PVP film is non-rigid in C H C 1 . For this particular film, the dry mass of oxidised copolymer (including counter ion required by electroneutrality) corresponds to deposition of 29 nmol of ferrocene sites. The deposition process involved passage of 3.32 mC of charge, i.e. 34 nmol of ferrocene sites were oxidized in total. This implies an (average) deposition efficiency for this experiment of 85%. If one were (erroneously) to assume a rigid film and calculate the extent of deposition by simple subtraction of the liquid contribution from the liquid+polymer data, the result would be ca. 9 nmol of ferrocene sites, i.e. factor of 3-4 in error. For comparison, analogous experiments with the P V F homopolymer showed that Sauerbrey equation-based calculations of coverage from the limiting frequency change during deposition underestimated the coverage by a factor of about two (30). Comparison of coverage estimates based on dry polymer coated electrode frequencies and frequency changes associated with redox cycling in water (ion and solvent exchange) are in excellent ( ± 1 0 % ) agreement. This indicates rigidity in aqueous media and air, but not in C H C 1 . 2

2

2

2

Conclusions Exchange of species between a solution and a polymer film is an established means of probing solution composition. The quartz crystal microbalance can monitor such exchange processes with high sensitivity. When combined with selectivity via electrochemical control and appropriate choice of polymer, the E Q C M becomes an attractive sensor. In order that the potential advantages of the E Q C M can be realised, certain criteria must be met. In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

12. HILLMAN ET AL.

Electrochemical Quartz Crystal Microbalance

161

Ν ACTIVITY CONTROL

Figure 4: Flow chart for using E Q C M data to distinguish electroneutrality and activity control under transient conditions from each other and from equilibrium. Absolute values of time-dependent quantities (i and M ) are employed. Firstly, when relating (gravimetrically determined) equilibrium film composition to solution composition, activity effects must be taken into account. These effects may cause solvent and other neutral molecules, as well as the target species, to enter/leave the film. The importance of medium effects cannot be overemphasised here. A special case is a co-ordination model, where favourable interaction between polymer and target species results in saturation of the film except at very low concentration. The film mass is then independent of solution composition. This situation is likely for systems where a strong, specific polymer/analyte interaction has been synthetically designed into the polymer. Secondly, selectivity is not always achievable. For example, permselectivity of ion-exchanging polymer films fails at high electrolyte concentration. We have shown that even if permselectivity is not thermodynamicaUy found, measurements on appropriate time scales in transient experiments can lead to kinetic permselectivity. To rationalise this behaviour we recall that the thermodynamic restraint, electrochemical potential, can be split into two components: the electrical and chemical terms. These conditions may be satisfied on different time scales. Dependent on the relative transfer rates of ions and net neutral species, transient responses may be under electroneutrality or activity control. Figure 4 shows how these cases may be distinguished on the basis of E Q C M data. (The equivalent plot for integrated data is obtained by replacing i with Q and M with ΔΜ in the ordinates of the plots.) In each decision box, one asks whether the designated plot (for a complete redox cycle) is hysteresisfree. We commonly find that both i vs. Ε and M vs. Q plots show hysteresis, indicating activity control. We suggest that neutral species transfers are slower than ion transfers, since the latter can be enhanced by migration. In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

162

BIOSENSORS AND CHEMICAL SENSORS

Finally, we rely on the Sauerbrey equation for the conversion of frequency data to mass data (and thus solution composition). This procedure is only valid for rigid polymer films. We therefore regard establishment of rigidity as vital. Acknowledgments We thank the S E R C (GR/E/32946), N A T O (86/0830) and the A i r Force Office of Scientific Research (87-0037) for financial support. M J S thanks the S E R C for a research studentship.

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

Murray, R.W. In ElectroanalyticalChemistry; Bard, A.J., Ed.; Marcel Dekker: New York, N Y 1984, Vol. 13; p.192. Szentirmay, M.R.; Martin, C.R. Anal. Chem. 1984, 56, 1898. Guadalupe, A.R.; Abruna, H.D. Anal. Chem., 1985, 57, 142. Nagy, G; Gerhardt, G.A.; Oke, A.F.; Rice, M.E.; Adams, R.N.; Moore, R.B.; Szentirmay, M.N.; Martin, C.R. J. Electroanal. Chem. 1985, 188, 85. Wier, L.M.; Guadalupe, A.R.; Abruna, H.D. Anal. Chem. 1985, 57, 2009. Ebersole, R.C.; Ward, M.D. J. Am. Chem. Soc., 1988, 110, 8623. Bruce, J.A.; Wrighton, M.S. J. Am. Chem. Soc. 1982, 104, 74. Bruckenstein, S.; Hillman, A.R. J. Phys. Chem., 1988, 92, 4837. Cox, J.A.; Kulesza, P.J. Anal. Chim. Acta, 1983, 154, 71. Sauerbrey, G.Z. Z. Phys., 1959, 155, 206. Applications of Piezoelectric Crystal Microbalances; Lu, C.; Czanderna, A.W., Eds.; Methods and Phenomena; Elsevier: New York, NY, 1984; Vol.7. Bruckenstein, S; Shay, M . Electrochim. Acta, 1985, 30, 1295. Kanazawa, K.K.; Gordon, J.G. Anal. Chim. Acta, 1985, 175, 99. Schumacher, R. Angew. Chemie (Int. Ed. English) 1990, 29, 329. Deakin, M.R.; Buttry, D.A. Anal. Chem. 1989, 61, 1147A. Buttry, D.A. In "Applications of the Quartz Crystal Microbalance to Electrochemistry"; Bard, A.J., Ed.; "Electroanalytical Chemistry"; Marcel Dekker Inc.: New York, NY, 1991, Vol. 17; pp.1-85. Kaufman, J.H.; Kanazawa, K.K.; Street, G.B. Phys. Rev. Lett. 1984, 53, 2461. Orata, D.; Buttry, D.A. J. Am. Chem. Soc. 1987, 109, 3574. Baker, C.K.; Reynolds, J.R. J.Electroanal.Chem. 1988, 251, 307. Reynolds, J.R.; Sundaresan, N.S.; Pomerantz, M.; Basak, S.; Baker, C.K. J.Electroanal.Chem. 1988, 250, 355. Baker, C.K.; Reynolds, J.R. Synth. Met. 1989, 28, C21. Hillman, A.R.; Loveday, D.C.; Bruckenstein, S. J. Electroanal. Chem., 1989, 274, 157. Hillman, A.R.; Loveday, D.C.; Swann, M.J.; Eales, R.M.; Hamnett, Α.; Higgins, S.J.; Bruckenstein, S.; Wilde, C.P. Faraday Disc. Chem. Soc. 1989, 88, 151.

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.

12. HILLMAN ET AL. 24. 25. 26.

Downloaded by IOWA STATE UNIV on October 18, 2014 | http://pubs.acs.org Publication Date: April 23, 1992 | doi: 10.1021/bk-1992-0487.ch012

27. 28. 29. 30. 31.

32. 33. 34. 35. 36.

Electrochemical Quartz Crystal Microbalance

163

Bruckenstein, S.; Wilde, C.P.; Shay, M . ; Hillman, A.R.J.Phys. Chem., 1990, 94, 787. Bruckenstein, S.; Wilde, C.P.; Hillman, A.R. J. Phys. Chem. 1990, 94, 6458. Hillman, A.R.; Swann, M.J.; Bruckenstein, S. J. Electroanal. Chem., 1990, 291, 147. Kelly, A.J.; Ohsaka, T.; Oyama, N.; Forster, R.J.; Vos, J.G.J.Electroanal. Chem. 1990, 287, 185. Inzelt, G. J. Electroanal. Chem. 1990, 287, 171. Naoi, K.; Lien, M.; Smyrl, W.H. J. Electrochem. Soc. 1991, 138, 440. Hillman, A.R.; Loveday, D.C.; Bruckenstein, S. Langmuir, 1991, 7, 191. Buck, R.P. Electrochemistry of Ion-Selective Electrodes In Comprehensive Treatise of Electrochemistry; White, R.E.; Bockris, J.O.; Conway, B.E.; Yeager, E., Eds.; Plenum Pub. Co.: New York, NY, 1984; Vol.8; pp 137248. Hillman, A.R.; Swann, M.J.; Bruckenstein, S. J. Phys. Chem. 1991, 95, 3271. Hillman, A.R.; Loveday, D.C.; Bruckenstein, S. J. Electroanal. Chem. 1991, 300, 67. Hillman, A.R.; Loveday, D.C.; Bruckenstein, S.; Wilde, C.P. J.C.S. Faraday Transactions, 1990, 86, 437. Bruckenstein, S.; Hillman, A.R.; Swann, M.J. J. Electrochem. Soc., 1990, 137, 1323. Borjas, R.; Buttry, D.A. J. Electroanal. Chem., 1990, 280, 73.

RECEIVED December 10, 1991

In Biosensors and Chemical Sensors; Edelman, P., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1992.